tkonolige commented on pull request #6671: URL: https://github.com/apache/incubator-tvm/pull/6671#issuecomment-716780203
Here are benchmark results on linux using forking. This is `dense` with 64 x 64 matrices. ntrials=16, 5 bencmarking iterations. | Branch | Auto-what | Target | Mean Time | STD | | ------ | --------- | ------ | --------- | --- | | main | autotvm | llvm | 20.429478 | 1.914881 | | PR | autotvm | llvm | 26.400752 | 9.013005 | | main | autotvm | cuda | 7.010600 | 0.080864 | | PR | autotvm | cuda | 7.001349 | 0.017357 | | main | autoscheduler | llvm | 16.711073 | 2.572820 | | PR | autoscheduler | llvm | 14.752735 | 0.757895 | | main | autoscheduler | cuda | 32.981713 | 0.072185 | | PR | autoscheduler | cuda | 32.283536 | 0.054209 | Looks pretty similar except for high standard deviation in runtime. Not sure what's causing this. I still think we should merge this as it fixes users not being able to run auto scheduling on Macs. ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org
